Groundwater contamination source identification based on Sobol Sequences-based Sparrow Search Algorithm with a BiLSTM surrogate model DOI Creative Commons

Yuanbo Ge,

Wenxi Lu, Zidong Pan

et al.

Research Square (Research Square), Journal Year: 2022, Volume and Issue: unknown

Published: Dec. 22, 2022

Abstract In the traditional linked simulation-optimization method, solving optimization model requires massive invoking of groundwater numerical simulation model, which causes a huge computational load. present study, surrogate origin was developed using Bidirectional Long and Short-term Memory neural network method (BiLSTM). Compared with models built by shallow learning methods (BP network) LSTM methods, BiLSTM has higher accuracy better generalization performance while reducing The to solved Sparrow Search Algorithm based on Sobol sequences (SSAS). SSAS enhances diversity initial population sparrows introducing introduces nonlinear inertia weights control search range efficiency. SSA, stronger global ability faster And identifies contamination source location release intensity stably reliably. This study also applied Cholesky decomposition establish Gaussian field for hydraulic conductivity evaluate feasibility method.

Language: Английский

Simultaneous identification of groundwater pollution source and important hydrogeological parameters considering the noise uncertainty of observational data DOI
Chengming Luo, Wenxi Lu, Zidong Pan

et al.

Environmental Science and Pollution Research, Journal Year: 2023, Volume and Issue: 30(35), P. 84267 - 84282

Published: June 26, 2023

Language: Английский

Citations

6
Simultaneous identification of a non-point contaminant source with Gaussian spatially distributed release and heterogeneous hydraulic conductivity in an aquifer using the LES-MDA method DOI
Wenjun Zhang, Teng Xu, Zi Chen

et al.

Journal of Hydrology, Journal Year: 2024, Volume and Issue: 630, P. 130745 - 130745

Published: Jan. 22, 2024

Language: Английский

Citations

2
Breaking the mold of simulation-optimization: Direct forward machine learning methods for groundwater contaminant source identification DOI
Chaoqi Wang, Zhi Dou, Yan Zhu

et al.

Journal of Hydrology, Journal Year: 2024, Volume and Issue: 642, P. 131759 - 131759

Published: Aug. 10, 2024

Language: Английский

Citations

2
An integrated modelling framework for multiple pollution source identification in surface water DOI
Xiaodong Liu, Xuneng Tong, Lei Wu

et al.

Journal of Environmental Management, Journal Year: 2023, Volume and Issue: 347, P. 119126 - 119126

Published: Sept. 29, 2023

Language: Английский

Citations

4
Groundwater contamination source-sink analysis based on random statistical method for a practical case DOI
Han Wang, Wenxi Lu

Stochastic Environmental Research and Risk Assessment, Journal Year: 2022, Volume and Issue: 36(12), P. 4157 - 4174

Published: June 18, 2022

Language: Английский

Citations

5
Prediction Performance Comparison of Risk Management and Control Mode in Regional Sites Based on Decision Tree and Neural Network DOI Creative Commons
Wenhui Zhu, Jun He, Hongzhen Zhang

et al.

Frontiers in Public Health, Journal Year: 2022, Volume and Issue: 10

Published: May 26, 2022

The traditional risk management and control mode (RMCM) in regional sites has the defects of low efficiency, high cost, lack systematism. Trying to resolve these explore application possibility machine learning, a characteristic dataset for RMCM was established. Three decision tree (DT) algorithms (CHAID, EXHAUSTIVE CHAID, CART) two artificial neural network (ANN) [back propagation (BP) radial basis function (RBF)] were implemented predict sites. results showed that aspects accuracy (ACC), precision (PRE), recall ratio (REC), F 1 value, CART–DT superior CHAID–DT (E-CHAID–DT); BP–ANN RBF–ANN. However, inferior ACC, PRE, REC, value. model is good at non-linear mapping, it flexible structure over-fitting. case study typical county demonstration area confirmed extensibility method, method great potential prediction future.

Language: Английский

Citations

3
Adjoint subordination to calculate backward travel time probability of pollutants in water with various velocity resolutions DOI Creative Commons
Yong Zhang, Graham E. Fogg, HongGuang Sun

et al.

Hydrology and earth system sciences, Journal Year: 2024, Volume and Issue: 28(1), P. 179 - 203

Published: Jan. 11, 2024

Abstract. Backward probabilities, such as the backward travel time probability density function for pollutants in natural aquifers/rivers, have been used by hydrologists decades water quality applications. Calculating these however, is challenging due to non-Fickian pollutant transport dynamics and velocity resolution variability at study sites. To address issues, we built an adjoint model deriving a backward-in-time fractional-derivative equation subordinated regional flow, developed Lagrangian solver, applied model/solver trace diverse flow systems. The subordinates reversed field, transforms forward-in-time boundaries into either absorbing or reflective boundaries, reverses tempered stable define mechanical dispersion. corresponding solver efficiently projects super-diffusive dispersion along streamlines. Field applications demonstrate subordination model's success with respect recovering release history, groundwater age, source locations various These include systems upscaled constant velocity, nonuniform divergent fields, fine-resolution velocities nonstationary, regional-scale aquifer, where significantly affects probabilities. Caution needed when identifying phase-sensitive (aqueous vs. absorbed) media. also explores possible extensions of quantifying probabilities more complex media, discrete fracture networks.

Language: Английский

Citations

0
Joint identification of contaminant source and dispersion coefficients based on multi-observed reconstruction and ensemble Kalman filtering DOI
Jing Li, Jun Kong, Mingjie Pan

et al.

Stochastic Environmental Research and Risk Assessment, Journal Year: 2024, Volume and Issue: 38(9), P. 3565 - 3585

Published: July 8, 2024

Language: Английский

Citations

0
Adjoint subordination to calculate backward travel time probability of pollutants in water with various velocity resolutions DOI Creative Commons
Yong Zhang, Graham E. Fogg, HongGuang Sun

et al.

Published: July 13, 2023

Abstract. Backward probabilities such as backward travel time probability density function for pollutants in natural aquifers/rivers had been used by hydrologists decades water-quality related applications. Reliable calculation of probabilities, however, has challenged non-Fickian pollutant transport dynamics and variability the resolution velocity at study sites. To address these two issues, we built an adjoint model deriving a backward-in-time fractional-derivative equation subordinated to regional flow, developed Lagrangian solver, applied model/solver backtrack various flow systems. The applies subordination reversed field, converts forward-in-time boundaries either absorbing or reflective boundaries, reverses tempered stable define mechanical dispersion. corresponding solver is computationally efficient projecting super-diffusive dispersion along streamlines. Field applications demonstrate that can successfully recover release history, dated groundwater age, spatial location(s) source(s) systems with upscaled constant velocity, non-uniform divergent fine-resolution velocities non-stationary, regional-scale aquifer, where significantly affects characteristics. Caution needed when identifying phase-sensitive (aqueous versus absorbed) source media. Possible extensions are also discussed tested quantifying more complex media, discrete fracture networks.

Language: Английский

Citations

0
Optimal design of groundwater pollution monitoring network based on a back-propagation neural network surrogate model and grey wolf optimizer algorithm under uncertainty DOI Creative Commons

Xinze Guo,

Jiannan Luo, Wenxi Lu

et al.

Research Square (Research Square), Journal Year: 2023, Volume and Issue: unknown

Published: July 31, 2023

Abstract In the optimal design of groundwater pollution monitoring network (GPMN), uncertainty simulation model always affects reliability when applying simulation–optimization methods. To address this issue, in present study, we focused on source intensity and hydraulic conductivity. particular, utilized Monte Carlo methods to determine layout scheme for wells under these conditions. However, there is often a substantial computational load incurred due multiple calls model. Hence, employed back-propagation neural (BPNN) develop surrogate model, which could substantially reduce load. We considered dynamic plume migration process GPMN. Consequently, formulated long-term GPMN optimization conditions with aim maximizing accuracy each period. The spatial moment method was used measure approximation degree between interpolated actual plume, effectively evaluate superior accuracy. Traditional easily trapped local optima solving so grey wolf optimizer (GWO) algorithm solve A hypothetical example designed evaluating effectiveness our method. results indicated that BPNN fit input–output relationship from as well significantly GWO solved improved solution distribution period be accurately characterized by optimized network. Thus, combining addressed problem uncertainty. developed stable reliable methodology optimally designing

Language: Английский

Citations

0